All cases

Review alerts at a Bengaluru lender

Bayes and default flags · Fictional case

The situation

Among 10,000 borrowers, 2% default. The model flags 75% of defaulters and 10% of other borrowers.

Your task

What fraction of flagged borrowers defaults?

Reveal step-by-step solution
  1. Set up the problem

    There are 200 defaulters and 9,800 other borrowers.

  2. Work through the numbers

    True alerts = 200 × 75% = 150. False alerts = 9,800 × 10% = 980. Total alerts = 1,130.

  3. Interpret the result

    Default probability given a flag is 150 ÷ 1,130 ≈ 13.3%. Use the flag to prioritize review.

What to remember

Build a natural-frequency table and put every flagged borrower in the denominator.

Review the topic cheat sheet